Metadata based combined approach for effective collaborative recommendation

Abstract

In this paper, we propose content-metadata based combined approach to effective collaborative recommendation. Our approach combines user-item rating scores and/or trust network information with content-metadata compensatively for boosting collaborative recommendation. In experiment, we identified that our approach could considerably improve recommendation performance when compared to existing collaborative recommendation methods.